Presented by Bill Dueber
Looked at specific users at a specific institution, but advocating that everyone should be doing this.
Who uses your library, and what are they trying to do?
But where do you draw the line? What is your library?
Why do we care?
- Not using? – it could be don’t like it, not
- Stopped using? – Google is the competition.
We’re not good at self-assessment, because nothing asks. No one cuts funding because not doing ‘some thing’ well enough. Incentive to find out if doing something bad is very low.
But you don’t want to be left by the wayside. We do good stuff and we want to be able to keep doing it.
Ask Your Reference Librarians
Because they’re the ones in the ‘trenches’ teaching and interacting with students and faculty.
Data-driven Decisions
The average patron interacts little with a librarian.
We need to make decisions based on the data that we have.
Unfortunately, APEs and PUGs hold undue sway over our decisions. Heat map shows only the search box, hours/ask us, but takes up very little of the front page.
After the redesign, looks more solid.
The Hard Truth
All the really hard problems are political. “My users” are more important than other users.
Electronic Users Need a Voice
You can be that advocate with the data that you have.
The easy, but important stuff: number of times view, when users are using it, device accessing, section patterns. Not great, but better than nothing. Google Analytics answers all the easy questions and more, but need to check privacy policy.
Finding ‘known-items’ covers 70% of the searches. Google effect to always search for it, not remember. Need to optimize for it.
Difficulty to do relevancy ranking, especially if in a different language.
Usability covers whether something is usable, not useful.
Not interested in clicks, interested in what people do. Differentiate between things with broad use and those with narrow but intense use.
Favourites actually being used by staff for something staff don’t have, but don’t care, because still looking for those electronic users.
Don’t recommend doing it this way, but want to record the data somehow: action, search text/index, record acted upon, numbers of results, index in list, link text/link target domain, sessionID, timestamp.
Might find some interesting data.
Do people do a search, then: write down the call number? can’t find what they want? not in the same building?
Made a very conscious decision not to fix prev/next/back buttons issue, because of the very low usage.
Don’t know why the numbers are the way they are.
What academic discipline are they looking at? (database of LC call numbers that match subjects) Will be interested to see how it changes over time.
Collection development application: look at what people are searching but not finding.
Takeaways
Give that voice to the electronic users.